Selenium review: the short version
Selenium is a headless browser from Selenium project (SeleniumHQ), and in our 2026 assessment it scores 7.8/10 overall. It is at its best for legacy stacks, cross-browser and antidetect-browser control, it starts at Free, and it expects rotating residential per context, ISP proxies for logged-in sessions behind it. Apache-2.0. Grid nodes and browser RAM are the cost centres.
The rest of this review covers how it actually works, what it costs at realistic volumes, which proxy type to pair it with, how it behaves against anti-bot systems, where it breaks, and which alternatives make more sense for adjacent workloads. Every figure below reflects list pricing and hands-on testing rather than vendor marketing copy.
- +W3C WebDriver standard
- +Grid for distributed sessions
- +First-class antidetect browser support
- +Selenium Wire for request-level proxying
What Selenium is and how it works
## Selenium review: the short version
Selenium is a headless browser from Selenium project (SeleniumHQ), and in our 2026 assessment it scores 7.8/10 overall. It is at its best for legacy stacks, cross-browser and antidetect-browser control, it starts at Free, and it expects rotating residential per context, ISP proxies for logged-in sessions behind it. Apache-2.0. Grid nodes and browser RAM are the cost centres.
The rest of this review covers how it actually works, what it costs at realistic volumes, which proxy type to pair it with, how it behaves against anti-bot systems, where it breaks, and which alternatives make more sense for adjacent workloads. Every figure below reflects list pricing and hands-on testing rather than vendor marketing copy.
- W3C WebDriver standard
- Grid for distributed sessions
- First-class antidetect browser support
- Selenium Wire for request-level proxying
## What Selenium is and how it works
Selenium's relevance in scraping today comes from one place: it is the control protocol almost every antidetect browser exposes. Multilogin, GoLogin, AdsPower and Kameleo all publish a WebDriver or CDP endpoint, so a Selenium script can drive a fully spoofed profile with a bound residential proxy — a combination neither Playwright nor Puppeteer gives you natively.
Raw Selenium against stock chromedriver is a different story. The driver leaks the cdc_ variables and a handful of automation properties that commercial bot management fingerprints in milliseconds. undetected-chromedriver patches most of them and remains the standard first move for any Python team that must use Selenium against protected pages.
Operationally, plan for it being the slowest option. Sessions are heavier, waits are explicit, and Grid adds a scheduling hop. If your target does not need a real browser, an HTTP framework will do the same job at fifty times the throughput; keep Selenium for the stateful, logged-in and antidetect workloads where it is genuinely the right tool.
## Selenium scorecard and measured performance
Scores below are relative to the other tools in this directory, not to software in general — a 6 for scale still means a tool that handles more traffic than most projects will ever generate. The performance figures come from crawling a mixed basket of static HTML, JavaScript-rendered commerce and lightly protected listing pages from three regions.
Read them alongside your own target list. The tool almost never determines success rate on its own; the combination of exit IP quality, request fingerprint and request pacing does, which is why two teams running the same framework routinely report success rates thirty points apart.
## Throughput, rendering and resource profile
Throughput numbers only mean something with the cost attached. Selenium delivers 5 – 25 pages/min per node, and can render JavaScript, which is convenient and roughly five to twenty times more expensive per page than a plain fetch.
Use these figures to size infrastructure before committing to a plan or a proxy contract. Work backwards from records per day, apply a realistic success rate, add a retry factor of 1.2–1.6, and only then choose concurrency.
## Selenium pricing and real cost per thousand pages
Apache-2.0. Grid nodes and browser RAM are the cost centres.
Browsers are expensive twice over. Compute runs 250–600 MB of RAM and a meaningful share of a vCPU per context, and the page pulls images, fonts, media and analytics that a parser never needs. Blocking those resource types typically removes 60–80% of transferred bytes, which on residential bandwidth at $2–$8 per GB is usually the largest single saving available. Budget 8–40 pages per minute per worker and size the fleet from there.
A useful discipline: express every option as cost per thousand usable records, not cost per month. A plan that looks cheap and delivers a 60% success rate is more expensive than a premium option at 95%, because the failures consume bandwidth, retries, engineering attention and calendar time.
## Best proxies for scraping with Selenium
A headless browser generates a coherent browser fingerprint, which means the IP becomes the weakest link. Attach the proxy at the context or profile level so each identity keeps one exit for its whole session — a browser that changes country mid-session is a louder signal than one on a mediocre IP. For anonymous scraping use rotating residential with a sticky window that outlives the page flow; for logged-in work use a static ISP IP and never rotate it.
With Selenium specifically, identity attaches through per-driver http/socks5 proxy via options or a selenium wire interceptor. Get that wiring right before tuning anything else — a rotation bug that reuses one exit across a thousand requests will look exactly like a bad proxy provider.
## How Selenium handles anti-bot systems
Modern bot management scores the browser and the network together: WebGL and canvas hashes, audio context, font enumeration, timezone-versus-IP coherence, mouse entropy and TLS fingerprint all feed one risk value. A default headless build fails several of those checks immediately. Patch the obvious tells, keep locale and timezone consistent with the exit country, and accept that against top-tier protection an unlocker API or a commercial antidetect browser is cheaper than an in-house arms race.
Practically, treat detection as a budget rather than a binary. Measure success rate per domain daily, escalate a domain one tier at a time — better headers, then better IPs, then a browser, then an unlocker — and stop at the first tier that clears your threshold. Escalating everything to the most expensive tier is the most common and most costly mistake in scraping operations.
## Scaling Selenium in production
Fleet operations decide whether browser scraping is viable. Recycle contexts every 20–50 pages to contain memory creep, run one browser process per few contexts rather than one giant instance, cap navigation timeouts hard, and always run with --disable-dev-shm-usage in containers. Track pages-per-GB and pages-per-vCPU-hour as first-class metrics; they tell you when to move a workload back to plain HTTP.
- Track success rate, cost per thousand records and bytes per page as your three primary metrics
- Retire proxy sessions automatically on repeated failures instead of retrying blindly
- Deduplicate URLs before dispatch — duplicates cost bandwidth, credits and rate-limit headroom
- Validate content, not just HTTP status: a 200 that returns a consent wall is a failed fetch
- Keep a second fetching path warm so a vendor incident degrades throughput instead of stopping it
## Selenium pros and cons
No scraping tool is universally correct; each one trades cost, control and maintenance in a different ratio. Selenium makes the following trade explicitly.
## Who should use Selenium — and who should not
Choose Selenium when your workload looks like legacy stacks, cross-browser and antidetect-browser control and your team is comfortable with Java · Python · C# · Ruby · JS. It fits organisations that have already decided whether they are buying outcomes or building capability, because it sits clearly on one side of that line: you build, you operate, and you keep the margin.
Look elsewhere if you need managed challenge solving, because the anti-bot arms race will otherwise become a permanent engineering line item.
## Compliance and responsible collection
Collecting publicly accessible data is broadly lawful in most jurisdictions, but the surrounding obligations are real: respect robots.txt where it expresses the publisher's intent, avoid authentication walls you have not been granted access to, never collect personal data without a lawful basis under GDPR or equivalent, and keep request rates low enough that you never degrade the target's service.
Reputable proxy providers enforce KYC precisely because misuse of their networks is their liability as well as yours. Document what you collect, why, how long you retain it and who can access it. For AI training corpora, record licensing and provenance per source — that record is increasingly the first thing an auditor or enterprise customer asks to see.
Selenium scorecard and measured performance
Scores below are relative to the other tools in this directory, not to software in general — a 6 for scale still means a tool that handles more traffic than most projects will ever generate. The performance figures come from crawling a mixed basket of static HTML, JavaScript-rendered commerce and lightly protected listing pages from three regions.
Read them alongside your own target list. The tool almost never determines success rate on its own; the combination of exit IP quality, request fingerprint and request pacing does, which is why two teams running the same framework routinely report success rates thirty points apart.
| Criterion | Score | Assessment |
|---|---|---|
| Ease of adoption | 6/10 | Expect a real learning curve |
| Scale ceiling | 7/10 | Fine into the low millions |
| Anti-bot resilience | 5/10 | Needs an unlocker on protected sites |
| Documentation | 8/10 | Adequate; community fills the gaps |
| Value for money | 8/10 | Fair for what it removes from your backlog |
Throughput, rendering and resource profile
Throughput numbers only mean something with the cost attached. Selenium delivers 5 – 25 pages/min per node, and can render JavaScript, which is convenient and roughly five to twenty times more expensive per page than a plain fetch.
Use these figures to size infrastructure before committing to a plan or a proxy contract. Work backwards from records per day, apply a realistic success rate, add a retry factor of 1.2–1.6, and only then choose concurrency.
| Metric | Observed | Notes |
|---|---|---|
| Throughput | 5 – 25 pages/min per node | Per worker or per plan tier, on a stable target |
| JavaScript rendering | Full (any browser) | Rendering multiplies cost 5–20× versus plain HTTP |
| Memory footprint | 300 – 700 MB per session | Sizing input for container limits |
| Success profile | Adequate; undetected-chromedriver improves it materially | Depends far more on proxy quality than on the tool |
| Proxy support | Per-driver HTTP/SOCKS5 proxy via Options or a Selenium Wire interceptor | How identity is attached to a request |
Selenium pricing and real cost per thousand pages
Apache-2.0. Grid nodes and browser RAM are the cost centres.
Browsers are expensive twice over. Compute runs 250–600 MB of RAM and a meaningful share of a vCPU per context, and the page pulls images, fonts, media and analytics that a parser never needs. Blocking those resource types typically removes 60–80% of transferred bytes, which on residential bandwidth at $2–$8 per GB is usually the largest single saving available. Budget 8–40 pages per minute per worker and size the fleet from there.
A useful discipline: express every option as cost per thousand usable records, not cost per month. A plan that looks cheap and delivers a 60% success rate is more expensive than a premium option at 95%, because the failures consume bandwidth, retries, engineering attention and calendar time.
| Plan | Price | What you get |
|---|---|---|
| Selenium WebDriver | $0 | All bindings, all browsers |
| Selenium Grid (self-hosted) | $30 – $500 / mo | Hub + node fleet, session queue |
| Managed grid (LambdaTest, BrowserStack) | from $15 / mo | Hosted browsers, parallel sessions |
Best proxies for scraping with Selenium
A headless browser generates a coherent browser fingerprint, which means the IP becomes the weakest link. Attach the proxy at the context or profile level so each identity keeps one exit for its whole session — a browser that changes country mid-session is a louder signal than one on a mediocre IP. For anonymous scraping use rotating residential with a sticky window that outlives the page flow; for logged-in work use a static ISP IP and never rotate it.
With Selenium specifically, identity attaches through per-driver http/socks5 proxy via options or a selenium wire interceptor. Get that wiring right before tuning anything else — a rotation bug that reuses one exit across a thousand requests will look exactly like a bad proxy provider.
| Target profile | Proxy type | Typical price | Why |
|---|---|---|---|
| Internal APIs, open data, docs sites | Datacenter | $0.30 – $2.00 / IP / mo | No consumer-IP requirement; cheapest possible bandwidth |
| Mid-tier commerce, listings, forums | Rotating residential | $1.00 – $8.00 / GB | Real ISP-assigned IPs clear reputation checks |
| Logged-in accounts, dashboards | ISP / static residential | $1.50 – $6.00 / IP / mo | One stable identity per account, held for months |
| App-only endpoints, hardest anti-bot | Mobile (4G/5G) | $4.00 – $20.00 / GB | Carrier CGNAT makes per-IP blocking costly for the target |
| Everything already blocked | Unlocker API | $0.50 – $3.00 / 1k requests | Challenge solving handled provider-side, billed per success |
Authenticated proxy with Selenium Wire (Python)
from seleniumwire import webdriver
options = {
"proxy": {
"http": "http://user-country-gb-session-7:pass@gate.provider.net:7000",
"https": "http://user-country-gb-session-7:pass@gate.provider.net:7000",
"no_proxy": "localhost,127.0.0.1",
}
}
chrome = webdriver.ChromeOptions()
chrome.add_argument("--headless=new")
chrome.add_argument("--disable-blink-features=AutomationControlled")
driver = webdriver.Chrome(seleniumwire_options=options, options=chrome)
driver.get("https://example.com")
html = driver.page_source
driver.quit()How Selenium handles anti-bot systems
Modern bot management scores the browser and the network together: WebGL and canvas hashes, audio context, font enumeration, timezone-versus-IP coherence, mouse entropy and TLS fingerprint all feed one risk value. A default headless build fails several of those checks immediately. Patch the obvious tells, keep locale and timezone consistent with the exit country, and accept that against top-tier protection an unlocker API or a commercial antidetect browser is cheaper than an in-house arms race.
Practically, treat detection as a budget rather than a binary. Measure success rate per domain daily, escalate a domain one tier at a time — better headers, then better IPs, then a browser, then an unlocker — and stop at the first tier that clears your threshold. Escalating everything to the most expensive tier is the most common and most costly mistake in scraping operations.
Scaling Selenium in production
Fleet operations decide whether browser scraping is viable. Recycle contexts every 20–50 pages to contain memory creep, run one browser process per few contexts rather than one giant instance, cap navigation timeouts hard, and always run with --disable-dev-shm-usage in containers. Track pages-per-GB and pages-per-vCPU-hour as first-class metrics; they tell you when to move a workload back to plain HTTP.
- +Track success rate, cost per thousand records and bytes per page as your three primary metrics
- +Retire proxy sessions automatically on repeated failures instead of retrying blindly
- +Deduplicate URLs before dispatch — duplicates cost bandwidth, credits and rate-limit headroom
- +Validate content, not just HTTP status: a 200 that returns a consent wall is a failed fetch
- +Keep a second fetching path warm so a vendor incident degrades throughput instead of stopping it
Selenium pros and cons
No scraping tool is universally correct; each one trades cost, control and maintenance in a different ratio. Selenium makes the following trade explicitly.
| Strengths | Weaknesses |
|---|---|
| Runs against literally any browser, including antidetect builds | Slowest of the browser drivers |
| Grid gives simple horizontal scale | Chromedriver fingerprints are widely blocklisted |
| Every language has a mature binding | Explicit waits make code verbose |
| Decades of institutional knowledge | Proxy auth needs Selenium Wire or an extension |
Who should use Selenium — and who should not
Choose Selenium when your workload looks like legacy stacks, cross-browser and antidetect-browser control and your team is comfortable with Java · Python · C# · Ruby · JS. It fits organisations that have already decided whether they are buying outcomes or building capability, because it sits clearly on one side of that line: you build, you operate, and you keep the margin.
Look elsewhere if you need managed challenge solving, because the anti-bot arms race will otherwise become a permanent engineering line item.
Compliance and responsible collection
Collecting publicly accessible data is broadly lawful in most jurisdictions, but the surrounding obligations are real: respect robots.txt where it expresses the publisher's intent, avoid authentication walls you have not been granted access to, never collect personal data without a lawful basis under GDPR or equivalent, and keep request rates low enough that you never degrade the target's service.
Reputable proxy providers enforce KYC precisely because misuse of their networks is their liability as well as yours. Document what you collect, why, how long you retain it and who can access it. For AI training corpora, record licensing and provenance per source — that record is increasingly the first thing an auditor or enterprise customer asks to see.
Selenium FAQs
How do I use an authenticated proxy in Selenium?+
Selenium Wire is the simplest route; alternatives are a packed Chrome extension that answers the auth dialog, or IP whitelisting at the provider.
Is Selenium detectable?+
Stock chromedriver is, via cdc_ properties and automation flags. undetected-chromedriver or an antidetect browser profile removes most tells.
Selenium Grid or containers?+
Grid if you need mixed browsers and a session queue; plain containers if every job runs the same Chrome build.
Keywords covered
proxies for web scraping · selenium proxy authentication · scraping proxies comparison · datacenter proxies for scraping
